DESCRIPTION: (Applicant's Description) The goal of this project is to elucidate the genetic and biological determinants of ovarian cancer, focusing on an in vitro model for ovarian cancer that we have developed (1-4). First, we propose to identify genes that are differentially expressed upon malignant transformation of HOSE cells using a modified suppression subtractive hybridization (SSH) approach. We will use consolidative SSH (CSSH) to identify genes that are either over-expressed or under-represented upon immortalization and malignant transformation of HOSE cells. Second, we propose to identify genes that are differentially expressed in HOSE cells heterozygous for BRCA1 (+/-) using CSSH. We have previously reported that the surface epithelial cells from ovaries of BRCA1 mutant allele carriers have a different phenotype in vitro compared to control surface epithelial cells and these features are present without the inactivation of the remaining wild-type allele of BRCA1. These results suggest that other genetic changes might precede loss of both copies of BRCA1 in the genesis of familial ovarian cancer. We will compare populations of HOSE cells for differentially expressed genes related to their BRCA1 carrier status. Third, we propose to determine the expression pattern of candidate ovarian cancer-causing genes identified by CSSH in benign and borderline ovarian tumors and ovarian carcinomas. To find the subset of genes discovered in our tissue culture model, which are relevant to clinical ovarian tumors, cDNA arrays representing the differentially expressed clones will be created. Reverse transcribed RNAs derived from benign, borderline, early and late stage ovarian tumors will be hybridized to individual arrays and the patterns of expression determined. Clones showing expression patterns consistent with the findings in the model will be evaluated for tissue specific expressions by multiple tissue northern blot analysis. Fourth, we propose to establish the pattern of expression of candidate ovarian cancer-causing genes in preneoplastic lesions by performing in situ hybridization on cancer prone ovaries. We have previously identified a preneoplastic phenotype in the ovaries from women at increased risk for developing ovarian cancer. Once we have confirmed the expression of several genes, in situ hybridization and immunohistochemical approaches will be used to evaluate their RNA and protein expression patterns at potentially the earliest stages of cancer development. Overall, the studies proposed should enable us to (i) determine if the histological phenotype we have previously described is likely to be a true precursor of ovarian cancer, (ii) establish whether other genetic changes precede BRCA1 inactivation in ovarian carcinogenesis, (iii) identify genes that are specifically expressed in the ovarian surface epithelium as opposed to the entire ovary, and (iv) identify early surrogate intermediate endpoint bio- markers of impending ovarian cancer.